The previous method (03_Build_GIF) to create a GIF in R from a series of images is surely the shortcut way, and we can do better.
The gganimate package can make this process much smoother, especially if we already have some understanding of the syntax in ggplot2.

Install then load the below packages.

library(transformr)
library(gganimate)

## Warning: package 'gganimate' was built under R version 3.5.2

## Warning: package 'ggplot2' was built under R version 3.5.2

library(ggplot2)
library(RColorBrewer)
library(dplyr)

## Warning: package 'dplyr' was built under R version 3.5.2

library(phytools)

## Warning: package 'ape' was built under R version 3.5.2

library(ggmap)

## Warning: package 'ggmap' was built under R version 3.5.2

Let’s read in the same data we’ve been playing with already.

goannas <- readRDS("Data/Goannas.RDS")
traits <- mutate(goannas$size.data,
                 TailRatio = Tail/Body_Length)

Before we get into plotting, we want to define a color scheme that we’re happy working with.

gramp <- colorRampPalette(brewer.pal(5, "YlGnBu"))
gcolors <- gramp(nrow(traits)); 
traits <- traits[order(traits$Body_Length),]
traits <- mutate(traits, plotcols = gcolors)

Now let’s plot the data, and animate it. Once we’ve got this GIF, I’ll try to explain how it’s all working.

TailSVL <- ggplot(traits, 
                  aes(x=log(Body_Length), 
                      y=log(Tail),
                      size=TailRatio)) +
  geom_point(alpha=0.75, 
             show.legend=FALSE, 
             color="black", 
             fill=traits$plotcols, 
             shape=21) + 
  # designate color here to keep the smoothness
  theme_bw() +
  scale_size(range=c(5,20)) +
  labs(title="Habitat: {closest_state}") + 
  # this will label the plot by the groupings
  transition_states(Habitat, 
                    transition_length=5, 
                    state_length=3) +
  ease_aes('cubic-in-out')

Save the GIF externally

anim_save(TailSVL, file="TailSVL.gif")

The beauty of gganimate is that we set up the plot just like we would our normal static ggplot figure.

ggplot(traits,
       aes(x=log(Body_Length),
           y=log(Tail),
           size=TailRatio)) +
  geom_point(alpha=0.75, 
             show.legend=FALSE, 
             color="black", 
             fill=traits$plotcols, 
             shape=21) + 
  # designate color here to keep the smoothness
  theme_bw() +
  scale_size(range=c(5,20))

We’ve plotted Tail length as a function of Body_Length, with each point colored scaled to the Tail/Body_Length ratio. Great.

All we need to animate this now is to split our plot up into a factor. This could be discrete bins of some value, or in our case, we’ll split the plots up by the preferred habitat type Habitat. We also choose how much time for the GIF to stay in each frame (state_length), and how long it should take to transition between frames (transition_length).

  transition_states(Habitat, 
                    transition_length=5, 
                    state_length=3) +
  ease_aes('cubic-in-out')

Sick one. Maybe we have time/data to create another!

How about animating the evolution of species ranges?

ranges <- filter(goannas$distribution.data, 
                 Name_in_Tree == "Varanus.mertensi" |
                 Name_in_Tree == "Varanus.giganteus")
ranges$Time <- 0

Grab the distribution for MRCA of these taxa (node 56)

mrcaVg <- filter(goannas$rase.data$DistData, 
                 Name_in_Tree == "n56")[,1:3]
mrcaVg$Name_in_Tree <- "Varanus.giganteus"

mrcaVm <- filter(goannas$rase.data$DistData,
                 Name_in_Tree == "n56")[,1:3]
mrcaVm$Name_in_Tree <- "Varanus.mertensi"

mrca <- rbind(mrcaVg, mrcaVm)
# set the time of the MRCA
mrca$Time <- round(max(nodeHeights(goannas$phy)) -
  nodeheight(goannas$phy, 56), 2)
# combine with the current distribution
ranges <- rbind(mrca, ranges)

Now we plot the ranges and animate the transitions
To recreate this, you’ll have to set up a google API in order to use ggmap and the function get_googlemap

rangemap <- get_googlemap(center = "Australia", zoom = 4, style = 'feature:all|element:labels|visibility:off')

## Source : https://maps.googleapis.com/maps/api/staticmap?center=Australia&zoom=4&size=640x640&scale=2&maptype=terrain&style=feature:all%7Celement:labels%7Cvisibility:off&key=xxx

## Source : https://maps.googleapis.com/maps/api/geocode/json?address=Australia&key=xxx

MRCA_range <- ggmap(rangemap) +
  stat_density2d(aes(x = Longitude, 
                     y = Latitude, 
                     fill = Name_in_Tree), 
                 alpha = 0.3, bins = 4,
                 geom = "polygon", data = ranges) + 
  theme(legend.position = 'none') +
  scale_fill_manual(values = c("#E31A1C", "#1F78B4")) +
  labs(title="Time: {closest_state} Million Years Ago") +
  transition_states(Time, 
                    transition_length=5, 
                    state_length=3) +
  ease_aes('cubic-in-out')

Save the GIF externally

anim_save(MRCA_range, file="N56_MRCA_range.gif")